JPH0229858B2 - - Google Patents

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention is directed to solving the problem of variation in fuel injection amount between cylinders of a fuel injection multi-cylinder internal combustion engine (hereinafter referred to as engine) such as a gasoline engine or a diesel engine. This invention relates to a fuel injection amount control method in which the amount of fuel is corrected for each cylinder based on the engine speed.

[Conventional technology]

Conventionally, fuel injection amount control for multi-cylinder engines has uniformly controlled the fuel injection amount for all cylinders, regardless of gasoline or diesel engine. That is, in the known electronically controlled fuel injection method for the Garrison engine, the opening time of the electromagnetic fuel injection valve disposed in each cylinder is controlled with the same control amount for all cylinders.
Furthermore, in electronically controlled diesel engines that have recently been put into practical use, injection amount control has been carried out by controlling the positions of control racks and spills, which are injection amount adjusting members common to the cylinders. For this reason, reducing the variation in injection amount between cylinders is done exclusively by strictly matching the characteristics of injection system parts (i.e., injection valves, injection pipes, etc.) in each cylinder. The current situation is that high manufacturing precision is required, increasing the cost.

Furthermore, even if the accuracy of the parts between the cylinders is raised to the limit, it is still completely powerless against changes over time and disturbances on the engine side, such as variations in the opening and closing timing of intake and exhaust valves, and as a result, all cylinders are the same. Stable fuel could not be obtained, and there was a high possibility that unpleasant periodic rotational fluctuations would occur, especially during idling.

In recent years, engine idle speeds have generally been kept low due to demands for improved fuel efficiency, and smoother idle speeds have been required for passenger cars in particular from the standpoint of comfort. A major issue for the time being is how to reduce periodic rotation fluctuations and achieve a low and stable idle.

To solve this problem, we detect variations in the rotational speed of each cylinder before and after fuel injection, and correct the fuel injection amount for each cylinder to equalize this variation in each cylinder, thereby adjusting the combustion state of each cylinder. A method for making it uniform has been proposed (Japanese Patent Application Laid-Open No. 58-214627).

Such correction control cannot avoid errors such as rounding errors due to loss of digits in digital calculations due to digital processing using a microcomputer or the like, or variations in characteristics of various sensors. Such an error occurs each time the injection amount of each cylinder is corrected, and as a result, the sum of the errors in the correction amount of each cylinder does not necessarily become zero. The sum of the errors can take either a positive or negative value, but since an error occurs independently for each correction, it is expected from the stochastic process that the sum of the errors will be accumulated sequentially through multiple corrections. . In fact, the sum of these correction errors sequentially accumulates and becomes a large absolute value, which affects the sum of the injection amount of each cylinder, that is, the injection amount of the entire engine, and the engine speed, emissions,
This may adversely affect drivability, engine performance, etc.

[Problem that the invention seeks to solve]

In view of the above points, the present invention provides a control device that corrects the fuel injection amount for each cylinder, which corrects the fuel injection amount for each cylinder without affecting the injection amount of the main body of the entire engine. The purpose is to provide a quantity control method.

[Means and actions for solving problems]

Therefore, in the present invention, we focus on the sum of the correction amounts for increasing and decreasing the fuel injection amount of each cylinder in order to equalize the generated torque of each cylinder, and we focus on the sum of the correction amounts of the fuel injection amount corrected for each cylinder. By increasing or decreasing the correction amount of the fuel injection amount so that it becomes zero or a value close to zero, the fuel injection amount for each cylinder can be corrected without affecting the original fuel injection amount, and all cylinders can be adjusted. The generated torque is made uniform, suppressing unpleasant vibrations in rotational speed, and ensuring stable rotation without gradual increase or decrease in rotational speed.

〔Example〕

Embodiments of the present invention will be specifically described below with reference to the drawings. FIG. 1 schematically shows the configuration of a four-cylinder diesel engine to which the present invention is applied. Publicly known 4
The cylindrical diesel engine 1 is equipped with, for example, a Botsushi VE type distribution injection pump 2 equipped with an electronic injection amount control device (so-called electronic governor). It is driven and rotated by an engine 1. Each cylinder of the engine 1 has an injection nozzle 3
1 to 34 are attached, and these nozzles 31 to 34 and the distribution injection pump 2 are connected to injection steel pipes 41 to 44.
The fuel pumped by the pump 2 at a predetermined timing is delivered to each of the nozzles 31 to 3.
4, a predetermined amount is injected into the combustion chamber (or auxiliary chamber) of each cylinder of the engine 1.

The pump drive shaft of the injection pump 2 has a third
As shown in the figure, 16
A disk 6 having six protrusions is provided, and a rotation speed sensor 5, which is, for example, a known electromagnetic pickup, is provided adjacent to the protrusions. Since the injection pump drive shaft rotates once every two rotations of the engine, eight pulses are output from the rotation speed sensor 5 every 45° crank angle, that is, per one rotation of the engine. Hereinafter, this signal will be referred to as the N signal and the explanation will proceed. This N signal is output as a rotational speed and constant crank angle progress signal to the control computer 9 shown in FIG. It receives the signal from the sensor 10 and calculates and determines the optimal fuel injection amount for the constantly changing engine operating conditions. In order to achieve the output injection amount, a drive signal is output to an injection amount control actuator 11 such as a linear solenoid attached to the injection pump 2.

Next, the detailed configuration of the distribution type injection pump 2 will be explained based on FIGS. 2 and 3. The base of the injection pump is a well-known Botsushi VE type injection pump, and the fuel suction, pressure feeding, distribution and injection timing control members and their operations are all well-known.
There is no difference from the VE type injection pump, so the explanation will be omitted. The feature of this pump is that the axial displacement of the spill ring 21, which is a fuel overflow metering member, is
It is controlled by an actuator 11 using a linear solenoid, and the injection amount is electronically controlled by a computer 9. When a control current output from the computer 9 is applied to the coil 23 of the actuator 11, a magnetic force of a strength corresponding to the control current is generated between the stator 24 and the moving core 25, and the moving core 25 is moved by the spring 30. It is struck by the reaction force and is pulled to the left in the figure. As the core 25 moves to the left, the lever 26, which is in contact with the core 25 at one end, rotates counterclockwise in the figure about the fulcrum 27 due to the tension of the spring 31. The lever 26 is connected to the spill ring 21 at the other end, and the spill ring 21 is moved to the right in the figure in conjunction with the above operation. In the VE type injection pump, as the spill ring 21 moves to the right in the figure, the overflow timing of fuel, that is, the end time of injection, is delayed, and as a result, the injection amount is increased. As explained above, if the current applied to the actuator 11 is increased, the injection amount will increase, and if the current is decreased, the injection amount will be decreased. Therefore, if the current applied to the actuator 11 is controlled by the computer 9, the injection amount can be controlled. It is.

In order to improve control accuracy, a position sensor 12 is installed coaxially with the actuator 11 in order to detect the actual position of the moving core 25 and correct the current flowing to the actuator 11 through position feedback control. , the position sensor 12 is integrally coaxial with the moving core 25 and includes a probe 28 made of ferrite or the like and a position detection coil 29.
It consists of Normal injection amount control is performed using the configurations shown in FIGS. 1 and 2 described above, and the computer 9 determines the optimum spill ring based on the N signal from the rotation speed detector 5 and the signal from the load sensor 10. The position, ie, the position of the moving core 25 of the actuator 11, is commanded, and the current applied to the actuator is controlled to obtain a target injection amount. however,
With only this basic injection amount, the injection amount is determined by the same common control amount for the four cylinders, so if the opening pressures of nozzles 31 to 34 vary, Naturally, the amount of fuel injected into the cylinders varies. In addition, as mentioned above, the N signal is the pump drive shaft 4.
A disk 6 with 16 protrusions integrally provided on the
A rotation sensor 5, which is, for example, a known electromagnetic pickup, is mounted on the pump housing so as to face the protrusion, and is output at every 45° crank angle of the engine.

In addition to the basic injection amount control described above, a calculation process within the computer 9 performs a correction process for injection amount variation between cylinders. First, the concept of control will be explained with reference to FIG. Figure 4 shows the N
The signal shows an example of a sequence chart of a known four-cylinder diesel engine.

Note that the shaded part on the sequence is the fuel injection timing to each cylinder.In the idling state where this control is mainly applied, the timing is usually several degrees of crank angle after top dead center. Fuel injection is performed. FIG. 4 shows the output obtained by processing the N signal in the computer 9, and shows the rotational fluctuation for each 45-degree crank angle of the engine. If we look closely at the relationship with the injection and explosion strokes of each cylinder, we can see that the N sensor output rises rapidly immediately after fuel injection and explosion, and then decreases as the next cylinder enters the compression stroke.

Therefore, the fine fluctuations in the N signal have a period of once every 1/2 rotation of the engine, and the maximum of the fluctuations is
It is experimentally known that the minimum value appears approximately every 90° crank angle of the engine. Here, if the difference between the maximum and minimum N fluctuation for each cylinder is ΔNi (i is the cylinder number in the combustion stroke at that time), then ΔNi is
It is known that there is a good correlation with the torque generated by combustion for each engine cylinder, and therefore,
If the ΔNi is made uniform across all cylinders #1 to #4, smooth idle rotation can be achieved.
Therefore, in this embodiment, the ΔN ₁ to ΔN ₄ are arithmetic averaged, that is, Δ= ₄ 〓 ⁱ⁼¹ ΔNi/4 is obtained, and the injection amount is controlled to increase or decrease so that the ΔNi of each cylinder is equal to the ΔN. . However, in the embodiment of the present invention, the N signal is simply output one after another at every 45° crank angle, so cylinder discrimination cannot be performed by comparing the combustion cycle of a specific cylinder as explained in FIG. (In order to achieve this, for example, one disk should be provided on the pump cam shaft, and one protrusion should be provided on the disk, for example, at a position that corresponds to the compression top dead center of the first cylinder.) In this example, even the cylinder discrimination is performed solely by software processing within the computer 9.

That is, among the four consecutively detected N signals,
The point in time when the minimum N signal is detected is determined to be the top dead center of one cylinder, that is, the position where Nmin is input in Fig. 4 is determined to be the top dead center, and each cylinder is identified sequentially thereafter. Cylinder discrimination is performed by

Next, the internal configuration of the computer 9 that executes the control concept described above and the actual processing executed within the computer 9 will be explained with reference to FIGS. 5 and 6. In FIG. 5, 100 is a microprocessor (MPU) that performs calculations for controlling the fuel injection amount.
Reference numeral 101 denotes the N signal counter, which counts the number of engine revolutions based on the N signal from the electromagnetic pickup 5. Moreover, this N signal counter 101 is
Interrupt control unit 102 in synchronization with engine rotation
An interrupt control signal is sent to each cylinder for each compression top dead center and 45° cam angle after top dead center.

Upon receiving this signal, the interrupt control section 102 outputs an interrupt signal to the microprocessor 100 via the common bus 150. Reference numeral 104 denotes an analog input port consisting of an analog multiplexer and an A/D converter, which has the function of A/D converting the accelerator opening degree, that is, the signal from the engine load sensor 10, and sequentially reading it into the microprocessor 100. Each of these units 101, 10
The output information of 2,104 is communicated to microprocessor 100 through common bus 150. 1
A power supply circuit 05 is connected to a battery 17 through a key switch 18 to supply power to the computer 9.

Reference numeral 107 denotes a temporary memory (RAM) which is used temporarily during program operation and into which stored contents can be sequentially written and read.In the RAM, rotation increments ΔN ₁ to ΔN ₄ for each combustion of the engine, which will be described later, are stored. , and correction values Q ₁ to _{Q 4} for correcting the control current to the fuel injection amount control actuator 11 for each combustion. A read-only memory (ROM) 108 stores programs, various constants, and the like.

109 is an output port that sets the control current to the actuator 11 calculated and determined by the MPU 100, and 110 is a drive circuit that converts the output signal into an actual operating current, which is connected to the linear solenoid actuator 11. There is. 11
1 is a timer that measures the elapsed time and
0, and also sends an interrupt control signal at regular intervals. As mentioned above, the N signal counter 101
counts the N signal and supplies two types of interrupt command signals to the interrupt control unit 102 for each compression top dead center of each cylinder of the engine and for each 45° cam angle after top dead center. The interrupt control unit 102 generates an interrupt signal from the signal, and causes the microprocessor 100 to execute an interrupt processing routine that will be explained below with reference to FIG.

FIG. 6 shows a flowchart of processing in the microprocessor 100. When the top dead center interrupt is started in step (201), first in step (202), 1 is added to the recognition number i value for recognizing which cylinder the current process will be performed on. Next, in step (203), it is checked whether the i value obtained by adding 1 is not 5, and if i=5, i is changed to 1 in step (204).
This is because this embodiment is disclosed for a four-cylinder engine. That is, if i=4,
This time, we will focus on the rotational fluctuations of the 4th cylinder, and in the next interrupt, i = 4 + 1 = 5, so we will rewrite it to i = 1 and perform the processing for the 1st cylinder again. That's why.

Next, in step (205), the instantaneous rotational speed N is read by the count of the N signal counter and the timer, and in step (206), the rotational speed N is stored as the current pre-injection rotational speed N _L i. Then, in step (207), the interrupt is terminated.

From then on, the microprocessor executes other processes other than this control until the next interrupt process, but in step (208), the microprocessor executes the second process every 45° cam angle after top dead center. When an interrupt command is received, interrupt processing for main control is started again. Step (209)
The instantaneous engine speed N at this time is detected by the N signal counter and timer as described above, and then in step (210), the engine speed N is stored as the current post-fuel injection speed N _H i. Then, in step (211), N _L i stored at the previous top dead center interrupt is subtracted from the N _H i, and this difference N _H
The ΔNi data in the RAM is rewritten with i−N _Li as the engine rotation increment ΔNi due to the current injection.
Next, in step (212), ΔN ₁ ~
The arithmetic mean Δ of N ₄ is calculated and the latest average rotation increment Δ is updated every time. That is, if the current processing is related to the fourth cylinder (i = 4), only ΔN ₄ is the new ΔN newly calculated from the old ΔN ₄ from one cycle before in step (211). ₄ and then calculate Δ, so Δ is always the average of the four most recent ΔNi.

At step (213), the difference d ² Ni=(Δ−ΔNi) between the above Δ and the current rotational increment ΔNi is calculated. From that value, in step (214),
Control current correction term Qj for each injection in RAM107
A variation correction integral amount ε=f(d ² Ni), which is a variable term for further correction, is determined. said ε and said
The relationship with d ² Ni is as shown in Figure 7, and the above
The correction integral amount ε takes a positive or negative value according to the value of d ² Ni. The value of the correction integral amount ε is stored in ROM1 prepared in advance.
It is obtained by searching the data within 08 from the value of d ² Ni. In step (215),
The value of ε is added to the value of Qi in the RAM 107.
It goes without saying that the value of Qi increases or decreases according to the positive or negative sign of the correction integral amount ε.

It is assumed that at the start of the control, Q ₁ =Q ₂ =Q ₃ =Q ₄ =0 is set by an initialization routine (not shown).

Next, in step (217), the absolute value of the current engine speed is calculated by averaging the N _L i and N _H i, and in step (218), the current engine load signal α is calculated from the analog input port 104. Enter. Then, in process 219, the basic control current Io of the actuator corresponding to the basic fuel injection amount to the cylinder where fuel will be injected next after processing the current is
For example, if the data in the ROM prepared in advance is
It is obtained by searching a two-dimensional map from .
In this control, the basic control current Io is corrected based on the magnitude of the rotation integral for each engine combustion, but in the processing up to step (219), ΔNi is corrected for the cylinders in which fuel has already been injected. Based on this, the correction term Qi for the relevant cylinder is updated. Therefore, the actuator control current I output at the end of this process is not the correction term Qi obtained in this process, but the correction term that has already been updated and stored three cycles ago for the cylinder to which fuel will be injected immediately next. must reflect the In step (220), processing 2 is performed for this purpose.
Add 1 to the recognition number i used up to 19 and use this as the correction term cylinder corresponding number j, and if j = 5, based on this j, the correction term Qj for j is,
It is read out from the RAM 107, added to Io in step (223), and outputted to the output port in order to displace the actuator 11 in preparation for the next injection.
When j=5 in step (221), j is set to 1 again in step (222), and the process also proceeds to step (223). i.e. step (219)
If the processing up to i=3 had been performed for the third cylinder, then the next fuel injection would be in the fourth cylinder, so the RAM
If i = ₄ , that is, steps up to step (219) have been performed for the fourth cylinder, then j = 4 + 1 = 5 → J = 1, the injection control current to the first cylinder for next injection is Q ₁
Modify based on.

By repeating the above process each time, the injection amount is gradually reduced for cylinders where the rotation increment per combustion is larger than the average, and conversely, the injection amount is gradually reduced for cylinders where the rotation increment per combustion is smaller than the average. The engine speed is gradually increased until finally an operating condition is reached in which all cylinders have equal rotational increments, that is, all cylinders have equal rotational torque.

However, the above control alone can cause various errors such as the least significant bit (LSB) error when reading the N signal in steps (205) and (209), and rounding errors due to digit loss during Δ calculation in step (212). Due to the error or because the correction integral amount ε obtained in step (214) is not a linear function of d ² Ni, the sum of the injection correction term Qi for each cylinder for the four cylinders S= ₄ 〓 ⁱ⁼¹ Qi is often not zero. and,
This sum S accumulates over time and always tends to decrease or increase. An increase in the absolute value S of the total sum S affects the injection amount of the entire engine, causing problems. In order to solve this problem, an interrupt processing routine explained according to the flowchart of FIG. 8 is executed.

At step (301), a real-time interrupt signal is received at regular intervals, and subsequent processing (hereinafter referred to as neutralization correction) is started.

Next, in step (302), the sum S=ΣQi of the injection amount correction terms Qi for each cylinder in the RAM 107 is calculated.

Next, in step (303), the absolute value |S| of the above-mentioned sum S is compared with a threshold value δ set to a value smaller than a value corresponding to a value that has an adverse effect on the engine (for example, 0.5 mm ³ /st). Determine size and absolute value |
If S| is smaller than the threshold value δ, the process ends without executing anything. If the absolute value |S| is larger than the threshold value δ, the process proceeds to the next step (304) in order to correct the injection amount correction term Qi.

In steps (304), (305), and (306), a fixed correction amount Δ is added to or subtracted from the injection amount correction term Qi of all cylinders according to the positive or negative sign of the sum S. As a result, the absolute value |S| of the sum S of the injection amount correction term Qi is corrected to decrease.

By repeating the above-mentioned processing at regular intervals, the absolute value |S| of the above-mentioned sum S can be determined by the threshold value δ
can be kept within.

Here, it is assumed that a real-time interrupt is performed in step (301), but it is not handled as an interrupt process.
Of course, it is also possible to use the timer 111 as a watchdog timer to constantly monitor the time and execute the routine following step (302) at regular intervals.

In reality, the real-time interrupt of step (301) is set every 10 mS, the threshold value δ of step (303) is set to a value equivalent to 0.2 mm ³ /st, and the constant correction amount Δ of steps (305) and (306) is set. As 0.05mm ³ /
Good results were obtained as a value equivalent to st.

If the time interval between real-time interrupts is too long,
The speed of neutralization correction by this process (the product of the above correction amount Δ and the neutralization correction rotation per unit time) is the learning correction speed of the injection correction amount Qi of each cylinder (the above ε and the number of explosion strokes per unit time). (product of
During the learning correction of Qi, the effect of the neutralization correction is not sufficiently exerted, and the sum S of the injection amount correction term Qi becomes a large value transiently, causing the above-mentioned problem.

Furthermore, if the one-time correction amount Δ of the neutralization correction is too large, temporary vibration of the engine will occur due to too large a discontinuous change in the injection amount due to the neutralization correction.

Therefore, it is desirable that the speed of the neutralization correction be equal to or slightly faster than the learning correction speed of the injection amount correction amount Qi.

The above neutralization correction is based on the injection correction term Qi for each cylinder.
to change by a certain amount at the same time, the total sum S
is corrected, but the amount for correcting the variation in injection amount for each cylinder is stored as the difference (Qi - Qk) between the injection correction terms Qi for each cylinder. Therefore, there is no need to impose any special conditions when performing the neutralization correction, and even if the neutralization correction is always performed, there will be no adverse effect on the engine.

In addition, when upper and lower limits are set for the injection amount correction amount for each cylinder, by performing the above neutralization correction, the sum of the injection amount correction amounts is kept close to zero, so the dynamic range of the injection amount correction is has the advantage of being constant.

[Other Examples]

In the first embodiment described above, the correction is performed periodically by a fixed amount Δ at fixed intervals. In this embodiment, neutralization correction is executed every time the injection correction amount for each cylinder is calculated.

That is, in the flowchart shown in FIG. 6, in step (215), the injection amount correction term for each cylinder is
After Qi is calculated, the neutralization correction process explained in the flowchart shown in FIG. 9 is executed, and then the process returns to the flowchart shown in FIG.
The following routine for determining the current value of the injection amount control actuator 11 is executed. That is, in the second embodiment, the steps shown in the flowchart of FIG. 9 are inserted between and in the flowchart of FIG. 6, and the injection amount is controlled.

Referring to FIG. 9, in step (402),
Total sum of injection quantity correction terms Qi in RAM107 = ₄ 〓 ⁱ⁼¹ Qi
is calculated.

Next, in step (403), the deviation per cylinder β=S/4 is calculated by dividing the above-mentioned sum S by the number of cylinders.

Next, in steps (404) and (405), the deviation β is subtracted from the injection quantity correction term Qi for all cylinders and stored in the RAM as a new injection quantity correction term Qi.
It goes without saying that the value of the injection amount correction term Qi is increased or decreased according to the positive or negative sign of the deviation β.

With this, the neutralization correction is completed, and next, the process moves to step (217) in FIG.

As a result, the sum S of the injection quantity correction terms Qi is always zero, and an equal and stable rotational torque is generated in all cylinders, resulting in an extremely stable and smooth operating state of the engine.

In the embodiment described above, an application example of the present invention to a fuel injection amount control device that detects the rotational speed of a four-cylinder diesel engine and corrects the injection amount of each cylinder has been described in detail. If the present invention is a control device that corrects the injection amount for each cylinder in a multi-cylinder internal combustion engine, it can also be applied to other control devices, such as a device that detects vibration acceleration before and after combustion in each cylinder and corrects the injection amount. Of course, it can also be applied in the same way.

The present invention uses a conventional control device that controls all cylinders in the same way, and can control the injection amount for each cylinder without any adverse effects on others, and the total injection amount for all cylinders is unchanged from the conventional one. Therefore, optimal control can be achieved by directly using the conventional adaptive values of the engine and control system obtained from many experiments (for example, the two-dimensional map of Io obtained from α when searched in step (219) above). It has the advantage of being able to do Another advantage is that the dynamic range of the correction amount for each cylinder can be made constant.

〔Effect of the invention〕

As described above, according to the present invention, the injection amount of each cylinder is corrected to equalize the generated torque for each cylinder,
Furthermore, since the sum of the correction values is set close to zero, there are no unpleasant periodic fluctuations in engine rotation, and stable and constant rotation can be maintained over a long period of time for smooth control. It has excellent effects.

[Brief explanation of drawings]

FIG. 1 is an overall configuration diagram showing an embodiment of the present invention;
Fig. 2 is a partial sectional view of the fuel injection pump in Fig. 1, Fig. 3 is a sectional view showing the installation state of the rotation speed sensor in Fig. 1, and Fig. 4 is an explanation of the operation of cylinder-specific injection amount correction. Fig. 5 is a timing explanatory diagram for the
The detailed configuration diagram of the control computer in the figure, FIGS. 6 and 8 are flowcharts showing the processing procedure in the control computer of this embodiment, FIG. 7 is a chart showing the variation correction integral amount, and FIG. 9 is the invention of the present invention. 3 is a flowchart showing a processing procedure in another embodiment. 1... Diesel engine, 2... Fuel injection pump, 5... Rotation speed sensor, 6... Disc, 9...
Control computer, 10...Load sensor, 11...
...Injection amount control actuator, 31, 32, 3
3, 34...Injection nozzle, 100...Microprocessor, 107...Temporary memory, 108...
...Read-only memory.

Claims (1)

[Claims] 1. A fuel injection amount control method for an internal combustion engine in which fuel is injected and supplied to a multi-cylinder internal combustion engine by a fuel injection device, in which the fuel injection amount of each cylinder is corrected to increase or decrease in order to equalize the generated torque for each cylinder. The injection amount control method is characterized in that each correction amount of the fuel injection amount is increased or decreased so that the sum of the correction amounts of the fuel injection amount corrected for each cylinder becomes zero or a value close to zero. Fuel injection amount control method.